A 3D Visualization Method for Weather Radar Data
By preprocessing and grid removal of weather radar data, combined with WebGL graphics card accelerated rendering, the problem of low computing and rendering performance in three-dimensional visualization of weather radar data is solved, and efficient cloud cluster shape retention and rapid analysis support is achieved.
Patent Information
- Application Number
- CN202410478889.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-20
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-04-20
AI Technical Summary
The three-dimensional visualization of existing weather radar data is time-consuming in computing performance and rendering performance, and is prone to excessive smooth distortion, and cloud cluster shape errors, especially when cloud clusters are dispersed, it is difficult to accurately analyze.
By preprocessing the weather radar echo reflectivity data, an initial three-dimensional grid is established, and the same value grid points are eliminated. Multi-data layer-by-layer rendering is adopted, and the rendering is accelerated using WebGL graphics card, using high numerical warm tones and low transparency , low numerical cool tone, high transparency strategy, hierarchically loading and rendering spatial grid points.
The amount of data involved in the calculation has been greatly reduced, avoiding the problem of smoothing out the shape of cloud clusters in surface analysis, improving the data loading speed, retaining the initial shape of cloud clusters, realizing effective visualization of cloud clusters, and supporting rapid meteorological development Business analysis.
Smart Images

Figure CN118244272B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of meteorological forecasting, and in particular to a three-dimensional visualization method for weather radar data. Background Art
[0002] Weather radar is a type of meteorological radar and is the main tool for monitoring and warning severe convective weather. Its working principle is to emit a series of pulsed electromagnetic waves and use the scattering and absorption of electromagnetic waves by precipitation particles such as clouds, rain, and snow to detect the spatial distribution and vertical structure of precipitation, and use this as a warning to track the precipitation system. When the electromagnetic waves emitted by the radar are projected onto cloud and rain particles, scattered electromagnetic waves are generated, among which the backscattered electromagnetic waves are received by the radar, which is the radar echo.
[0003] Weather radar needs to scan the atmosphere at various altitudes (to be precise, by adjusting the elevation angle of the radar scan) to obtain the physical conditions of each layer of the atmosphere. Therefore, the data of weather radar is composed of multiple layers of volume scan data, with about 400,000 data points in each layer. According to 20 layers of data, there are about 8 million points. Displaying 8 million points in three-dimensional space is very time-consuming for computing performance and visualization rendering performance. The existing spatial three-dimensional surface analysis has a huge amount of running volume, and after rendering, it is too smooth and distorted. When the clouds are scattered, it is easy to analyze the wrong shape of the cloud. This is all caused by the huge amount of weather radar data. Summary of the invention
[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a three-dimensional visualization method for weather radar data, which greatly reduces the amount of data involved in the calculation while ensuring the business needs of meteorological analysis.
[0005] To achieve the above object, the present invention can be implemented through the following technical solutions:
[0006] A three-dimensional visualization method for weather radar data comprises the following steps:
[0007] Get the weather radar echo reflectivity data value;
[0008] Preprocessing weather radar echo reflectivity data values;
[0009] Establishing an initial three-dimensional grid with the pre-processed weather radar echo reflectivity data values;
[0010] The data inside the grid points with the same value that are clustered together are removed;
[0011] Establish sequence output for grid points with the same value;
[0012] Load and render spatial grid points hierarchically based on sequence output.
[0013] Further, the preprocessing of the weather radar echo reflectivity data values includes the following steps:
[0014] Obtain the weather radar echo reflectivity data values between 5 dBZ and 75 dBZ;
[0015] Perform interval sampling on the weather radar echo reflectivity data values, and divide them into one scale value every 5 dBZ;
[0016] The divided scale values are used as the grid point data values of the initial three-dimensional grid.
[0017] Further, the preprocessing of the weather radar echo reflectivity data values includes the following steps:
[0018] Classify the weather radar echo reflectivity data values according to the risk level;
[0019] The echo reflectivity data values of the first level are G1, where 0 dBZ < G1 ≤ 22 dBZ; the echo reflectivity data values of the second level are G2, where 22 dBZ < G2 ≤ 35 dBZ; the echo reflectivity data values of the third level are G3, where 35 dBZ < G3 ≤ 50 dBZ; the echo reflectivity data values of the fourth level are G4, where G4 > 50 dBZ; the echo reflectivity data values of each level are used as the grid point data values of the initial three-dimensional grid.
[0020] Further, the step of establishing the initial three-dimensional grid with the preprocessed weather radar echo reflectivity data values includes:
[0021] Use the altitude returned by the radar as the first dimension Z, the longitude as the second dimension X, and the latitude as the third dimension Y;
[0022] Use the preprocessed weather radar echo reflectivity data values as the values of the three-dimensional grid points;
[0023] Establish the initial three-dimensional grid.
[0024] Further, the step of removing the internal data of the co-valued grid points gathered together includes the following steps:
[0025] Locate a grid point;
[0026] Judge whether the value of this grid point is the same as that of its surrounding grid points;
[0027] When the values of the surrounding grid points are all the same as that of this grid point, it is considered that this grid point is surrounded by the same value, and this grid point is set to be empty;
[0028] When the values of the surrounding grid points are not all the same as that of this grid point, it is considered that this grid point is not surrounded by the same value, and the value of this grid point remains unchanged;
[0029] Use the same steps to traverse all grid points.
[0030] Furthermore, the points around each grid point are quickly located and indexed by adding 1 or subtracting 1 to the X, Y, and Z coordinate values.
[0031] Furthermore, the step of establishing the sequence output for the grid points with the same value comprises the following steps:
[0032] The set of identical values V = {V1, V2, ... V m},
[0033] Any grid points with the same value V use a two-dimensional array to store grid points with the same value: V = [[X1, Y1],..., [X m , Y m ]], where X is the longitude coordinate and Y is the latitude coordinate, thereby establishing a sequence output for grid points with the same value.
[0034] Furthermore, when the sequence output of the grid points with the same value is established, each sequence of the grid points with the same value is saved separately.
[0035] Furthermore, the step of loading and rendering the spatial grid points hierarchically according to the sequence output includes:
[0036] According to the sequence output, according to the value of V, the grid sequence with large values is loaded first for rendering. The strategy of high-value warm tones and low transparency and low-value cold tones and high transparency is adopted to load and render the spatial grid points in layers according to the values.
[0037] Furthermore, based on WebGL, the graphics card is used to accelerate the rendering to achieve a layered rendering effect with multiple layers.
[0038] Compared with the prior art, the present invention has the following beneficial effects:
[0039] According to the actual needs of meteorological services, the present invention eliminates data points that are not of concern to the business, which can save a large number of three-dimensional grid points, greatly reduce the amount of data involved in the calculation, and at the same time ensure the business needs of meteorological analysis. The present invention adopts a multi-data layer layer-by-layer rendering method to well restore the effect of cloud clusters and fog, and can give priority to rendering key numerical layers, providing better support for rapid business analysis. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 It is a schematic diagram of a method for rapid positioning and indexing of surrounding grid points of the present invention;
[0041] Figure 2 It is a rendering effect diagram of three-dimensional visualization of weather radar data of the present invention. DETAILED DESCRIPTION
[0042] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation methods. The terms such as "upper", "inner", "middle", "left", "right" and "one" cited in this specification are only for the convenience of description and are not used to limit the scope of the present invention. The change or adjustment of their relative relationship shall also be regarded as the scope of the present invention without substantially changing the technical content.
[0043] The multi-layer weather radar data can be regarded as a dense cloud, which is called the initial three-dimensional grid of radar data.
[0044] Embodiment 1
[0045] The three-dimensional visualization method of weather radar data of the present invention comprises the following steps:
[0046] Step S1, obtaining weather radar echo reflectivity data values of meteorological targets through weather radar.
[0047] Step S2, preprocessing the weather radar echo reflection data value, wherein the specific steps of preprocessing the weather radar echo reflectivity data value are as follows:
[0048] S21. Obtain weather radar echo reflectivity data values between 5 DBZ and 75 DBZ from the data returned by the radar. The echo reflectivity (DBZ) data values of the weather radar in the business range from 5 to 75 have analysis value. Therefore, the weather radar echo reflectivity data values between 5 and 75 are selected as initial data for processing.
[0049] S22, taking values of the weather radar echo reflectivity data interval, with each interval of 5 DBZ divided into a scale value;
[0050] S23. The divided scale values are used as grid point data values of the initial three-dimensional grid, so that the values of each grid point in the entire initial three-dimensional grid are multiples of 5 between 5 and 75. After processing, the data values of the entire three-dimensional grid have at most 15 (15=(75-5) / 5+1) values. The reduction in the number of value types makes it easier to determine adjacent grids with the same value. It is only necessary to determine whether the values of adjacent grids are equal.
[0051] Step S3: Establishing an initial three-dimensional grid with the pre-processed weather radar echo reflectivity data values. The specific steps are as follows:
[0052] S31, the altitude of the meteorological target returned by the radar is taken as the first dimension Z, the longitude is taken as the second dimension X, and the latitude is taken as the third dimension Y;
[0053] S32, using the pre-processed weather radar echo reflectivity data value as the value of the three-dimensional grid point;
[0054] S33. Establish an initial three-dimensional grid.
[0055] Step S4: remove the internal data of the grid points with the same value that are clustered together. Grid points with the same value that are clustered together can be regarded as solid clouds of the same color. In the visualization process, the internal data of the solid cloud is consistent with the edge data and does not need to be actually displayed. Therefore, the internal data of the solid cloud can be removed by edge detection, and the changes in the surrounding values can be judged based on any point in space. The specific steps are as follows:
[0056] S41, locating a grid point, for example, any grid point A in space;
[0057] S42, determining whether the values of the grid point and its surrounding grid points are the same;
[0058] S43, when the surroundings of the grid point (9 adjacent grid points in the upper layer, 9 adjacent grid points in the lower layer, and 8 surrounding grid points in the same layer) have the same value, it is considered that the point is surrounded by the same value, and the grid point is set to empty; since the grid point values have been processed into multiples of 5 between 5 and 75, the same surrounding values can be directly judged by whether the values are equal;
[0059] S44. When not all surrounding grid points have the same value as the grid point, it is considered that the grid point is not surrounded by the same value, and the value of the grid point remains unchanged;
[0060] S45, using the same rule as above to determine the value of the next grid point, the points around each grid point can be quickly located by adding 1 or subtracting 1 to the X, Y, and Z coordinate values, thereby speeding up the determination;
[0061] Use the same rule as above to traverse all grid points until all grid points in the space are traversed once and it is confirmed that all internal solid points have been removed. This method only requires one traversal.
[0062] After this step, the data in the initial three-dimensional grid has eliminated the wrapped grid points with the same value. According to the actual data situation most of the time, about 2 / 3 of the grid points have been reduced. In normal weather processes, such as cloudy and overcast days when the cloud reflectivity is relatively consistent, about 80% of the grid point calculations can be saved, greatly reducing the amount of data involved in the calculation.
[0063] Step S5, establishing a sequence output for grid points with the same value, including the following steps:
[0064] The set of identical values V = {V1, V2, ... V m},
[0065] Any grid points with the same value V use a two-dimensional array to store grid points with the same value: V = [[X1, Y1],..., [X m, Y m ]], where X is the longitude coordinate and Y is the latitude coordinate, so as to establish a sequence output for the grid points with the same value;
[0066] Each sequence of grid points with the same value can be saved separately for hierarchical loading in the visualization stage.
[0067] Step S6: Load and render the spatial grid points hierarchically according to the numerical values according to the sequence output. To achieve a better visualization analysis effect, load the points in a transparency and color gradient manner from small to large values. The lower the value, the colder the tone and the higher the transparency. The higher the value, the warmer the tone and the lower the transparency. According to the sequence output, according to the numerical value of V, the grid sequence with a large value is loaded first for rendering. The strategy of high-value warm tones and low transparency and low-value cold tones and high transparency is adopted to load and render the spatial grid points hierarchically according to the numerical values, as shown in the following table:
[0068]
[0069] Based on WebGL, graphics card accelerated rendering is used to achieve multi-layered rendering effects, which can realize fast loading and rendering layer by layer, and layered cloud atomization effects.
[0070] Embodiment 2
[0071] The three-dimensional visualization method of weather radar data of the present invention comprises the following steps:
[0072] Step S1, obtaining weather radar echo reflectivity data values of meteorological targets through weather radar.
[0073] Step S2, preprocessing the weather radar echo reflection data value, wherein the specific steps of preprocessing the weather radar echo reflectivity data value are as follows:
[0074] S21. Classify the weather radar echo reflectivity data values according to risk levels, for example: level 1 (no rain), level 2 (light rain), level 3 (moderate rain), and level 4 (heavy rain). This classification is more in line with our actual application scenarios and is easier to judge the weather conditions of meteorological targets. At the same time, it greatly reduces the number of data values of the entire three-dimensional grid (only four types), greatly reduces the amount of data calculation, and makes it easier to judge adjacent grids with the same value;
[0075] S22. The echo reflectivity data value of the first level is G1, where 0 dBZ < G1 ≤ 22 dBZ; the echo reflectivity data value of the second level is G2, where 22 dBZ < G2 ≤ 35 dBZ; the echo reflectivity data value of the third level is G3, where 35 dBZ < G3 ≤ 50 dBZ, and the echo reflectivity data value of the fourth level is G4, where G4 > 50 dBZ. It is not limited to these four levels and the division of such echo reflectivity data values, and there can be other division methods according to needs.
[0076] S23. The echo reflectivity data value of each level is used as the grid point data value of the initial three-dimensional grid.
[0077] Step S3. Establish an initial three-dimensional grid with the preprocessed weather radar echo reflectivity data value. The specific steps are as follows:
[0078] S31. Take the height of the meteorological target returned by the radar as the first dimension Z, the longitude as the second dimension X, and the latitude as the third dimension Y.
[0079] S32. Take the preprocessed weather radar echo reflectivity data value as the value of the three-dimensional grid point.
[0080] S33. Establish an initial three-dimensional grid.
[0081] Step S4. For the grid points with the same value gathered together,剔除 their internal data. The grid points with the same value gathered together can be regarded as a solid cloud mass of the same color. In the process of visualization, the internal data of the solid cloud mass does not need to be actually displayed because it is the same as the edge data. Therefore, the internal data of the solid cloud mass can be kicked out by the edge detection method, and the change of the surrounding values is judged based on any point in space. The specific steps are as follows:
[0082] S41. Locate a grid point, for example, any grid point A in space.
[0083] S42. Judge whether the value of this grid point is the same as that of its surrounding grid points.
[0084] S43. When all the surrounding grid points (9 adjacent grid points in the upper layer, 9 adjacent grid points in the lower layer, and 8 surrounding grid points in the same layer) of this grid point have the same value, it is considered that this point is surrounded by the same value, and this grid point is set to be empty.
[0085] S44. When the surrounding grid points are not all the same as the value of this grid point, it is considered that this grid point is not surrounded by the same value, and the value of this grid point remains unchanged.
[0086] S45. Judge the value of the next grid point with the same rule as above. The points surrounding each grid point can be quickly located and indexed by adding 1 or subtracting 1 to the X, Y, and Z coordinate values to speed up the judgment speed.
[0087] Use the same rule as above to traverse all grid points until all grid points in the space are traversed once and it is confirmed that all internal solid points have been removed. This method only requires one traversal.
[0088] After this step, the data in the initial three-dimensional grid has eliminated the wrapped grid points with the same value. According to the actual data of most of the time, about 75% of the grid points have been reduced. The lower the level, the more grid points are reduced. In normal weather processes, such as cloudy and overcast days when the cloud reflectivity is relatively consistent, about 85% of the grid point calculations can be saved, greatly reducing the amount of data involved in the calculation.
[0089] Step S5, establishing a sequence output for grid points with the same value, including the following steps:
[0090] The set of identical values V = {V1, V2, ... V m},
[0091] Any grid points with the same value V use a two-dimensional array to store grid points with the same value: V = [[X1, Y1],..., [X m , Y m ]], where X is the longitude coordinate and Y is the latitude coordinate, so as to establish a sequence output for the grid points with the same value;
[0092] Each sequence of grid points with the same value can be saved separately for hierarchical loading in the visualization stage.
[0093] Step S6: Load and render the spatial grid points hierarchically according to the numerical values according to the sequence output. To achieve a better visualization analysis effect, load the points in a transparency and color gradient manner from small to large values. The lower the value, the colder the tone and the higher the transparency. The higher the value, the warmer the tone and the lower the transparency. According to the sequence output, according to the numerical value of V, the grid sequence with a large value is loaded first for rendering. The strategy of high-value warm tones and low transparency and low-value cold tones and high transparency is adopted to load and render the spatial grid points hierarchically according to the numerical values, as shown in the following table:
[0094]
[0095] Based on WebGL, graphics card accelerated rendering is used to achieve multi-layered rendering effects, which can realize fast loading and rendering layer by layer, and layered cloud atomization effects.
[0096] The present invention eliminates data points that are not of concern to the business, greatly reducing the amount of data involved in the calculation, while ensuring the business needs of meteorological analysis. The present invention avoids smoothing out part of the cloud shape for surface construction in methods such as surface analysis and search. On the basis of improving the data loading speed, the initial shape of the cloud is retained without smoothing and denoising, which is convenient for the meteorological business to meet the needs of visual analysis of the rapid changes of small cloud clusters in severe convective weather processes. The present invention uses transparent and different coloring methods to retain the key information of each small cloud cluster, and also uses a multi-data layer layer-by-layer rendering method to well restore the effect of cloud cluster fog. The key numerical layer can be rendered first, providing better support for the rapid development of business analysis.
[0097] The implementation methods of the present invention are not limited to this. According to the above content of the present invention, using common technical knowledge and customary means in the field, without departing from the above basic technical ideas of the present invention, the present invention can also make other various forms of modification, replacement or combination, all of which fall within the scope of protection of the present invention.
Claims
1. A three-dimensional visualization method for weather radar data, characterized in that, it includes the following steps: Obtain the weather radar echo reflectivity data value; Preprocess the weather radar echo reflectivity data value; Establish an initial three-dimensional grid with the preprocessed weather radar echo reflectivity data value; For the grid points with the same value gathered together, eliminate their internal data; Establish a sequential output for the grid points with the same value; Load and render the spatial grid points hierarchically according to the sequential output; The step of eliminating the internal data of the grid points with the same value gathered together includes the following steps: Locate a grid point; Judge whether the value of this grid point is the same as that of its surrounding grid points; When the values of the surrounding grid points are all the same as that of this grid point, it is considered that this grid point is surrounded by the same value, and this grid point is set to be empty; When the values of the surrounding grid points are not all the same as that of this grid point, it is considered that this grid point is not surrounded by the same value, and the value of this grid point remains unchanged; Traverse all grid points with the same steps.
2. The three-dimensional visualization method for weather radar data according to claim 1, characterized in that, the step of preprocessing the weather radar echo reflectivity data value includes the following steps: Obtain the weather radar echo reflectivity data value between 5 dBZ and 75 dBZ; Take interval values for the weather radar echo reflectivity data value, and divide it into a scale value every 5 dBZ; The divided scale values are used as the grid point data values of the initial three-dimensional grid.
3. The three-dimensional visualization method for weather radar data according to claim 1, characterized in that, the step of preprocessing the weather radar echo reflectivity data value includes the following steps: Classify the weather radar echo reflectivity data value according to the risk level; The echo reflectivity data value of the first level is G1, 0 dBZ < G1 ≤ 22 dBZ; the echo reflectivity data value of the second level is G2, 22 dBZ < G2 ≤ 35 dBZ; the echo reflectivity data value of the third level is G3, 35 dBZ < G3 ≤ 50 dBZ; the echo reflectivity data value of the fourth level is G4, G4 > 50 dBZ; The echo reflectivity data value of each level is used as the grid point data value of the initial three-dimensional grid.
4. The three-dimensional visualization method for weather radar data according to claim 2 or 3, characterized in that, the step of establishing an initial three-dimensional grid with the preprocessed weather radar echo reflectivity data value includes: Take the height returned by the radar as the first dimension Z, the longitude as the second dimension X, and the latitude as the third dimension Y; Use the preprocessed weather radar echo reflectivity data value as the value of the three-dimensional grid point; Establish an initial three-dimensional grid.
5. The three-dimensional visualization method for weather radar data according to claim 1, characterized in that, The points around each grid point are quickly located and indexed by adding 1 or subtracting 1 to the X, Y, and Z coordinate values.
6. The three-dimensional visualization method for weather radar data according to claim 1, characterized in that, the step of establishing a sequential output for the grid points with the same value includes the following steps: The set of identical values V = {V 1 ,V 2 ,...V m }, Any grid point with the same value V uses a two-dimensional array to store grid points with the same value: V = [[X 1 , Y 1 ],...,[X m , Y m ]], where X is the longitude coordinate and Y is the latitude coordinate, thereby establishing a sequence output for grid points with the same value.
7. The three-dimensional visualization method for weather radar data according to claim 6, characterized in that: When the sequence output of the grid points with the same value is established, each sequence of the grid points with the same value is saved separately.
8. The method for three-dimensional visualization of weather radar data according to claim 7, It is characterized in that The step of loading and rendering spatial grid points hierarchically according to the sequence output comprises: According to the sequence output, according to the value of V, the grid sequence with large values is loaded first for rendering. The strategy of high-value warm tones and low transparency and low-value cold tones and high transparency is adopted to load and render the spatial grid points in layers according to the values.
9. The method for three-dimensional visualization of weather radar data according to claim 8, It is characterized in that Based on WebGL, graphics card accelerated rendering is used to achieve multi-layered rendering effects.
